基于自产热和外传热的锂离子电池热学模型参数辨识方法

Accurate identification of thermal model parameters (heat capacity and thermal resistance) of lithium ion batteries is crucial for thermoelectric coupling modeling and state parameter estimation of batteries. There is a certain conversion formula between the thermal model parameters and the thermophysical parameters (specific heat capacity and thermal conductivity) of lithium ion batteries, so the two sets of parameters have an equivalent relationship. The study of thermal model parameters can refer to the research methods of thermophysical parameters. Accurate measurement of battery specific heat capacity and thermal conductivity requires the use of expensive testing equipment, which is costly and takes longer to test. Using optimization algorithms to identify thermal model parameters not only has low cost, but also has short calculation cycle. In this paper, a fast method for identifying the thermal model parameters of NCM laminated soft packaged lithium-ion batteries is proposed based on the charging and discharging conditions combined with the battery heat generation and heat transfer mechanisms. By establishing a distributed thermal equivalent circuit model to simulate the temperature distribution along the thickness direction of the battery, the reversible heat can be ignored and the calculation of battery heat generation can be simplified by designing a bidirectional pulse operating condition; The temperature was discretized and the thermal model parameters of 0℃, 10℃, and 20℃ were identified using (adaptive particle swarm optimization (APSO) algorithm. The identified parameters are equivalently treated as specific heat capacity and thermal conductivity, with an average specific heat capacity of 0.996 J/(g·K) and an average thermal conductivity of 0.376 W/(m·K). Comparing the specific heat capacity and thermal conductivity of batteries from different literature, they are within a reasonable range, and the difference is not more than 8%. Temperature sensitivity analysis was conducted. When the thermal model parameters changed by ±5%, the temperature error of the model simulation results was less than 0.08℃. Therefore, in the range of 0 to 20℃, the thermal model parameters are not sensitive to temperature. Finally, the identified parameters are brought into the model and simulated using a different strategy than the parameter identification experiment. The measured and simulated battery surface temperatures agree well with a temperature error of less than 0.1℃, which proves that the proposed thermal model parameter identification method has a high accuracy. Moreover, the proposed method has the advantage of simplicity and simplicity, requiring only two batteries of the same specification, without the need for other devices, and having a short test cycle. It can provide technical support for the identification of thermal model parameters for stacked soft packaged lithium ion batteries.